arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1903.00519cs.LGcs.AIstat.ML
收藏

聚合解释方法以实现稳定和鲁棒的可解释性

Aggregating explanation methods for stable and robust explainability

DTU Compute, Technical University Denmark(丹麦技术大学计算学院)

Laura Rieger, Lars Kai Hansen

摘要:

尽管关于解释神经网络的文献越来越多,但在如何解释神经网络决策或如何评估解释方面尚未达成共识。我们在本文中的贡献是双重的。首先,我们调查的计划,结合解释方法,减少模型的不确定性,以获得一个单一的聚合解释。我们提供的证据表明,聚合是更好地确定重要的功能,比单独的方法。对解释的对抗攻击是最近一个活跃的研究课题。作为我们的第二个贡献,我们提出的证据表明,总的解释是更强大的攻击比个人的解释方法。

英文摘要:

Despite a growing literature on explaining neural networks, no consensus has been reached on how to explain a neural network decision or how to evaluate an explanation. Our contributions in this paper are twofold. First, we investigate schemes to combine explanation methods and reduce model uncertainty to obtain a single aggregated explanation. We provide evidence that the aggregation is better at identifying important features, than on individual methods. Adversarial attacks on explanations is a recent active research topic. As our second contribution, we present evidence that aggregate explanations are much more robust to attacks than individual explanation methods.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑